LOL you need help with an MMORPG lol don''t we all :D
Anyways here are the best answers i can give!

1)Not sure about a c++ book you saying you''ve never used c++ before? If so you should really hold off a little bit on a game in general and learn the basics and how it works etc.. (and how to print messages and such). Also tho if you want a free 3d engine for your RPG me and my friends are using www.genesis3d.com and if you want a 3d game i suggest you use it(though the commands took my friend a while to learn!). Also remember you''ll have to get a book for winsock(if your using winsock) and if you do DON''T GET ON BY MICROSOFT! I''m seroius i made that mistake and now my server doesn''t work and i have to pay another $50 to get another books. I''m going to try a networking for dummies book as those are usually helpful. If your using directplay, well can''t help you with that.

2)I don''t understand what you mean by "what kind of server." Please elaborate.

3)I really need to know this also!

4)I don''t know how they save their stats but I''m pretty sure (not totally now) that UO player-run servers use regular files but i could be wrong. I''m using SQL for mine.

5)It really depends on what your doing. You doing it for fun but then again people will come to your game, realize a lot of stuff is client side, and will have fun themselves hacking the program which really doesn''t make it fun for you! But what you siad was basically that all info stays on server and almost nothign on client (because graphics HAVE to be on the client ) Here''s what a proffesional game developer (who''s helping us out :p aren''t i special lol ) told me. "If you have everything client side then it''s easily hacked but your server goes fast. If you have everything server side it''s very hard to hack but you loose server speed. You have to find the perfect balance for your game." And so should you

6)What, do you mean collision detecting server side? Well we''re doing it client side because genesis3d woudl be fine client side and they''re so way (without running the engine) we could test it server side!

Also another thing you should think about ASAP or very soon. You should think if it''s going to be a pay game. I know it''s for fun (same for mine) but if it becomes popular then you''ll need to upgrade your connection, and if you wait too long noone will pay (a game once was free they won''t suddenly pay). That is why we''re making it a pay game. but do whatever you want just a suggestion you should think about.

1) One that college students learn from... I swear by "Object Oriented Programming in C++" by Waite Group Press. It covers a great deal of C++ programming, and newer releases should have development environment notes too.

2) Plan a multi-server architecture, running on an NT based OS, that heavily relies on multithreading. (Despite nay-sayers) it''s a very stable OS.

3) Run each client in a thread, that only does stuff based on incoming events. DO NOT check each client each loop, that''s just a waste of CPU time.

4) POL (a UO server program) saves data in text files. Head over to http://vulpin.burdell.org/pol/ and download and check out their program. Try running your own server to understand how each part of it works with the rest of the game. (This point is particularly important, I believe... nothing beats experience).

5) Clients should only play ''mother may I'' with the server. Let the server do all the work. This is important for large player-base games (2000-3000 as you mentioned) because of hackers that ruin the game for others.

6) Certainly, just learn fast/dirty ways to do it. It doesn''t need to be impressive or perfect.

7) Certainly.
- People WILL pay for something that has been free for a long time, that you suddenly give a price tag to. (I have proof of this, I work for a large company that only recently started selling their product that had been free for years.)
- You also need to listen to player feedback on those sorts of things. If you want to add a feature to the game, consult the players first. Always remember that they are your customer base, not your lab guinea pigs.
- Don''t try to overwhelm yourself the first time that you write a game. This sounds like your first venture into game programming. Make a small roam-the-world online game first, then add features to it as you see fit. Design your game engine before programming it, and design it modular.

GOod luck!


MatrixCubed
http://MatrixCubed.org

If you do collision detection server-side only then you will have alot of problems. People will be walking through walls on the client and get de-syched on the server. Unless your going to wait for server-side confirmation before moving the player (which will be very sluggish unless you have less than 50ms or so latency) then you need to do collision detection client-side and server-side. That was, the server will keep track of where your client is and your client wont end up walking through walls or going through impassible objects.

Um, guys... You rarely seem to know how to create even the most basic complete game, some of you are just learning C++ and you don''t seem to have an idea of basic network / multiplayer programming... And you want to create an MMORPG ??? c''mon, that''s a joke... Better do something you can really complete. This way you''ll have more fun and learn more... Believe me.

Tim

--------------------------
glvelocity.gamedev.net
www.gamedev.net/hosted/glvelocity

I use server side collision detection and it doesn''t add any lag. It''s just a matter of how you write it. The server completes the entire move checking for collisions before sending it to the client. This way the server can''t be outrun and "who was there first" isn''t an issue.

You should start simple and then add features and improve code as you go.

Just start with being able to connect more than one person and be able to walk around the world.

Ben
http://therabbithole.redback.inficad.com

quote:Original post by MatrixCubed
2) Plan a multi-server architecture, running on an NT based OS, that heavily relies on multithreading. (Despite nay-sayers) it''s a very stable OS.

I will have to agree with you here... NT with service pack 6 is rock solid. Of course, NT and Windows in general is very dependent on the hardware.

On NT here''s the equation:
poor hardware = poor performance;
good hardware = good performance;
great hardware = great performance;

You get the ideal




Game On,
David[Dak LozarLoeserElysian Productions, Inc